Evaluation of the failure rates of transmission lines during hurricanes using a neuro-fuzzy system

This paper proposes a method to evaluate the impact of extreme weather on the failure rates of transmission lines. The method is based on a neuro-fuzzy system: adaptive neuro-fuzzy inference system (ANFIS). ANFIS is a popular neuro-fuzzy system and it has the configuration of an artificial neural network (ANN) and functions as a fuzzy inference system (FIS). Actually, ANFIS uses an ANN to realize the function of a Sugeno-type FIS (S-FIS). In this way, ANFIS combines the merits of ANN and FIS. It has the learning feature of ANN and retains the interpretability of FIS. Therefore, ANFIS can use the learning ability of ANN to improve the performance of FIS. It is noted that the focus of this paper is the computation of failure rates to be used in reliability calculations in the hurricane environment. The proposed method is demonstrated by using the IEEE reliability test system (RTS).